maxjhandsome
/
llvm-project
forked from OSchip/llvm-project
1
0
Fork 0
Code Issues Pull Requests Packages Releases Wiki Activity

[X86] Rename hasNoSignedComparisonUses to hasNoSignFlagUses. Add the instruction that only modify the O flag to the waiver list. 

The only caller of this turns CMP with 0 into TEST. CMP with 0 and TEST both set OF to 0 so we should have no issues with instructions that only use OF.

Though I don't think there's any reason we would read just OF after a compare with 0 anyway. So this probably isn't an observable change.

llvm-svn: 349223
This commit is contained in:
Craig Topper 2018-12-15 01:07:19 +00:00
parent 5c304eac41
commit 1fc257d97f
1 changed files with 14 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

22
llvm/lib/Target/X86/X86ISelDAGToDAG.cpp
View File

@ -2192,8 +2192,8 @@ bool X86DAGToDAGISel::isSExtAbsoluteSymbolRef(unsigned Width, SDNode *N) const {
}
/// Test whether the given X86ISD::CMP node has any uses which require the SF
/// or OF bits to be accurate.
static bool hasNoSignedComparisonUses(SDValue Flags) {
/// flag to be accurate.
static bool hasNoSignFlagUses(SDValue Flags) {
// Examine each user of the node.
for (SDNode::use_iterator UI = Flags->use_begin(), UE = Flags->use_end();
UI != UE; ++UI) {
@ -2215,11 +2215,14 @@ static bool hasNoSignedComparisonUses(SDValue Flags) {
switch (FlagUI->getMachineOpcode()) {
// These comparisons don't treat the most significant bit specially.
case X86::SETAr: case X86::SETAEr: case X86::SETBr: case X86::SETBEr:
case X86::SETEr: case X86::SETNEr: case X86::SETPr: case X86::SETNPr:
case X86::SETEr: case X86::SETNEr: case X86::SETOr: case X86::SETNOr:
case X86::SETPr: case X86::SETNPr:
case X86::SETAm: case X86::SETAEm: case X86::SETBm: case X86::SETBEm:
case X86::SETEm: case X86::SETNEm: case X86::SETPm: case X86::SETNPm:
case X86::SETEm: case X86::SETNEm: case X86::SETOm: case X86::SETNOm:
case X86::SETPm: case X86::SETNPm:
case X86::JA_1: case X86::JAE_1: case X86::JB_1: case X86::JBE_1:
case X86::JE_1: case X86::JNE_1: case X86::JP_1: case X86::JNP_1:
case X86::JE_1: case X86::JNE_1: case X86::JO_1: case X86::JNO_1:
case X86::JP_1: case X86::JNP_1:
case X86::CMOVA16rr: case X86::CMOVA16rm:
case X86::CMOVA32rr: case X86::CMOVA32rm:
case X86::CMOVA64rr: case X86::CMOVA64rm:
@ -2241,6 +2244,9 @@ static bool hasNoSignedComparisonUses(SDValue Flags) {
case X86::CMOVNP16rr: case X86::CMOVNP16rm:
case X86::CMOVNP32rr: case X86::CMOVNP32rm:
case X86::CMOVNP64rr: case X86::CMOVNP64rm:
case X86::CMOVO16rr: case X86::CMOVO16rm:
case X86::CMOVO32rr: case X86::CMOVO32rm:
case X86::CMOVO64rr: case X86::CMOVO64rm:
case X86::CMOVP16rr: case X86::CMOVP16rm:
case X86::CMOVP32rr: case X86::CMOVP32rm:
case X86::CMOVP64rr: case X86::CMOVP64rm:
@ -3715,7 +3721,7 @@ void X86DAGToDAGISel::Select(SDNode *Node) {
if (isUInt<8>(Mask) &&
(!(Mask & 0x80) || CmpVT == MVT::i8 ||
hasNoSignedComparisonUses(SDValue(Node, 0)))) {
hasNoSignFlagUses(SDValue(Node, 0)))) {
// For example, convert "testl %eax, $8" to "testb %al, $8"
VT = MVT::i8;
SubRegOp = X86::sub_8bit;
@ -3723,7 +3729,7 @@ void X86DAGToDAGISel::Select(SDNode *Node) {
MOpc = X86::TEST8mi;
} else if (OptForMinSize && isUInt<16>(Mask) &&
(!(Mask & 0x8000) || CmpVT == MVT::i16 ||
hasNoSignedComparisonUses(SDValue(Node, 0)))) {
hasNoSignFlagUses(SDValue(Node, 0)))) {
// For example, "testl %eax, $32776" to "testw %ax, $32776".
// NOTE: We only want to form TESTW instructions if optimizing for
// min size. Otherwise we only save one byte and possibly get a length
@ -3738,7 +3744,7 @@ void X86DAGToDAGISel::Select(SDNode *Node) {
// be sure we calculate the correct sign flag if needed.
(CmpVT != MVT::i16 || !(Mask & 0x8000))) ||
CmpVT == MVT::i32 ||
hasNoSignedComparisonUses(SDValue(Node, 0)))) {
hasNoSignFlagUses(SDValue(Node, 0)))) {
// For example, "testq %rax, $268468232" to "testl %eax, $268468232".
// NOTE: We only want to run that transform if N0 is 32 or 64 bits.
// Otherwize, we find ourselves in a position where we have to do